Prion diseases or transmissible spongiform encephalopathies (TSEs) are infectious neurodegenerative conditions characterized by vacuolar degeneration of the central nervous system (CNS) and deposition of an abnormal isoform of the host-encoded prion protein (PrP). These diseases affect a wide variety of animal species and display limited zoonotic potential. Usually displaying prolonged incubation periods, they are clinically recognized via progressive neurological deterioration resulting from synaptic and neuronal loss and associated activated glial responses to CNS damage. Identification of genetically inherited forms of these diseases implicate further the critical role of the prion protein in disease pathogenesis and their classification as protein-misfolding disorders, with similarities to other progressive dementias such as Alzheimer’s and Parkinson’s diseases and amyotrophic lateral sclerosis. Sporadic prion diseases, with no known genetic risk factor or exposure to infection, have also been identified. The disease-associated isoform of the prion protein gains several properties including ability to transmit infection, limited protease resistance, and increased ability to fibrillize and form amyloid, these observations on both etiology and biochemical nature of the agent resulted in the prion hypothesis [1].

The prion hypothesis proposed that the infectious agent may be solely composed of a proteinaceous particle, i.e., the disease-associated isoform of the prion protein (PrP^Sc), with the means to self-propagate via an auto-catalytic process of template-mediated refolding of the nascent cellular prion protein (PrP^C). The pathway and mechanisms from refolding of the prion protein to neurodegeneration are still unknown. The peripheral pathogenesis of these diseases have been extensively studied using animal models, as mice are naturally susceptible to both sheep scrapie and bovine spongiform encephalopathy (BSE). Following natural peripheral exposure to prion agents, infection is usually sequestered to lymphoid organs prior to invasion of the nervous system (termed neuroinvasion) [2,3]. In the absence of local draining lymphoid tissue, subsequent circulation to other lymphoid organs and neuroinvasion are ultimately blocked [4,5]. Using transgenic mouse models or bone marrow chimeric mouse models various hypotheses regarding the genes and cells involved in the prion infectious pathway have been proven or refuted. Though often contradictory, results from these studies have revealed that disease-associated PrP is deposited in lymphoid follicles and replicates upon follicular dendritic cells (FDCs). This stage of pathogenesis has been shown to be dependent upon expression of cellular prion protein by FDCs [6,7]. The lymphoid stage of prion pathogenesis is not always obligatory and is dependent upon various factors including: the host, strain of infectious agent, route of infection and infectious dose [8,9,10,11]. Experimental delivery of prion agent directly to the nervous system may facilitate direct neuroinvasion [12,13]. Direct delivery of the prion agent to the central nervous system may bypass peripheral pathogenesis completely resulting in infection in naturally resistant hosts [14,15]. Concurrent lymphoid sequestration of infection likely still occurs after direct CNS infection and may play a role in subsequent pathogenesis [15,16].

Originating from the stromal cell compartment FDC are specialized cells that capture and retain antigen for presentation to cognate B-cells and subsequent generation of specific antibody responses and thus reside within primary B-cell follicles and germinal centers within lymphoid tissues [17,18]. Due to the reliance of FDCs on lymphotoxin signaling from B-cells to maintain FDC homeostasis [19], immune-compromised models lacking mature FDCs reveal deficits in peripheral prion pathogenesis. As such a positive relationship has been firmly established between the immune-competence of the host and ability to support prion pathogenesis. As with most infectious agents, typical pathogenesis occurs via the preferred route often as a compromise between the agent and the host. Prevention or blocking of this route may promote adaption or alternative pathways of pathogenic activity. For example blocking of intestinal Peyer’s patch formation prior to oral prion infection led to pathogenesis via FDC contained within isolated lymphoid follicles [4]. Similar altered pathogenic routes may be induced following splenectomy or sympathectomy [3,20]. Data obtained from prion pathogenesis studies reveal that rarely infection is blocked completely and more often disease is attenuated, revealing alterations to well-definable characteristics such as the disease incubation period or targeting of pathological changes both peripherally and within the host CNS. The difficulties in analyzing such data are compounded by the problems of global host models (i.e., gene-knockout) that effect both peripheral and CNS innate or adaptive immune responses. Despite these difficulties several clear messages have been received regarding the role of specific components of the innate immune system, some of which are less inter-dependent than the systems regulating the adaptive immune response, which we review and summarize below.

The innate immune system is considered to be a protective system that is older in evolutionary terms than the adaptive immune system. This system was originally proposed to respond in a non-specific way to invading pathogens in an attempt to eliminate them or their activity, whilst alerting or informing the adaptive immune response. As such the innate immune system confers no immunological memory or lasting protective immunity. The identification and characterization of specific pattern recognition receptors, such as Toll-like (TLR), C-type lectin (CLR), NOD-like (NLR) and RIG-I-like (RLR) receptors, has revealed the diversity of the innate immune system to detect and tailor specific responses to pathogen-associated molecular patterns (PAMPs) [21,22,23,24]. In line with the innate immune system constituting the first line of active defense after physical barrier mechanisms, a significant portion of the cellular component is pre-localized to potential exposure sites. This tissue-resident, or homeostatically maintained, population primarily assesses and responds to insults accordingly, and likely deals effectively with acute insults. A large systemic pool of both cellular (inflammatory) and proteinaceous components of the innate immune system also exists, to be mobilized rapidly in response to signals from epithelia or resident innate immune cells. Below we consider various components of the innate immune system, both cellular and proteinaceous, and their possible roles in prion pathogenesis (summarized in Figure 1).

The epithelial cell layers constitute the first physical barrier to infection. Prion infection via the skin is experimentally facilitated by scarification or breaking of the keratinized epidermal barrier layers to allow access to the epithelial layer beneath [25]. Prion infection via the oral or nasal route appears to be naturally more efficient. Prion pathogenesis can be greatly enhanced by disruption of the epithelial layers involved in these routes [26,27]. Under steady state conditions oral uptake of prion infection occurs via specialized microfold cells (M-cells) localized to the follicle-associated epithelium (FAE) of intestinal Peyer’s patches [28,29]. The developmental maturation of M-cells has been shown to require CD137 signaling, with CD137 ligand provided by hematopoietic cells (either B-cells or dendritic cells) [30]. To provide this signal hematopoietic cells are likely required to localize to the sub-epithelial vicinity via chemotaxis mediated by C-C chemokine receptor type 6 (CCR6) responding to C–C chemokine ligand 20 (CCL20), as M-cells do not form if CCR6-CCL20 signaling is prevented [31,32]. The PrP^C expressed by M-cells has been hypothesized to function as an uptake receptor for pathogenic bacteria [33]. Translocated PrP^Sc has also been observed on enterocyte derived extracellular vesicles suggesting an M-cell independent pathway for prions to cross epithelial barriers [34]. Lumenal sampling by dendritic cells (DC) has been observed [35]. Although this activity by classical DC is rare under steady-state conditions it suggests an additional potential prion uptake mechanism. The de-differentiation of M-cells prior to oral scrapie infection was sufficient to block pathogenesis completely [29], indicating that prion uptake via epithelial and DC mechanisms are insufficient to establish infection in the steady state. The incidence of DC sampling increased following epithelial TLR signaling [36]. Thus the possibility of coincident bacterial infection or other inflammatory stimuli enhancing prion uptake to threshold levels via this mechanism are worthy of further investigation.

Following uptake and translocation across the epithelial barrier layer, prions are then free to encounter other elements of the innate immune system. Directly beneath the FAE and within basolateral pockets of the M-cells themselves are areas enriched in classical DC, macrophages (Mϕ) and lymphocytes termed the subepithelial dome (SED), precisely located so as to process incoming antigen. The final physical barrier to prion pathogenesis appears to be the distance between site of entry and FDC and subsequently between FDC and peripheral nerves. Innervation of the lymphoid organs has been shown to be critical for prion neuroinvasion [20]. Manipulation to shorten the distance between FDC and peripheral nerves has shown to increase the rate of neuroinvasion [37]. These data provide a strong argument against hematogenous spread of prion agent directly to the CNS. Following aerosol exposure, prion pathogenesis occurs independently of the immune system, FDC, B cells, T cells, NK cells, lymphotoxin β receptor or CD40 ligand signaling [11].

The complement system is composed of numerous blood borne proteins that are generally synthesized in the liver and circulate as inactive precursors or pro-proteins. Significant amounts of complement components are also synthesized locally by tissue-resident mononuclear phagocytes (MNPs) [38,39], epithelial and stromal cells such as FDC [40]. The complement system is activated by numerous triggers that establish a proteolytic cleavage cascade, amplifying the response and typically resulting in the formation of the membrane attack complex (MAC). The classical complement activation pathway is triggered mainly by antigen-antibody complexes. Alternative activation pathways exist wherein complement component C3b may bind directly to foreign material (alternative pathway) or where mannose-binding lectin recognizes and binds to sugar moieties (lectin pathway) both associated with pathogenic microorganisms. The MAC functions to form transmembrane pores, thereby punching holes in pathogenic cells with the aim of promoting cell lysis and death.

Complement has been shown to be activated early during prion pathogenesis by as yet undetermined mechanisms and may constitute the first active response to infection. Prion protein has been shown to be directly bound by C1q and Factor H [41] and this binding occurs specifically when prion protein is conformationally modified to represent the conversion to the disease-associated isoform [42]. The role of complement in prion pathogenesis has recently been subject to review [43]. In brief prion or TSE agents are opsonized by complement components including C1q and C3, most likely via the classical complement activation pathway, which may aid in their targeting of the agent to lymphoid follicles. Mice lacking in complement components C1qa, C2 or C3 revealed deficient peripheral prion pathogenesis under specific conditions [44,45].

The local production of complement components by MNP has been shown to alter their function [46]. C1q enhances the receptor-mediated uptake of disease-associated PrP by classical dendritic cells (see section 2.4.2 below) [47]. The association of C1q with PrP^Sc may alter the downstream processing by classical DC, indeed Flores-Langarica and colleagues imply the receptor calreticulin in this process suggesting targeting towards a protein degradative pathway. This is counter-intuitive to the concept that classical DC may retain intact antigen via a FcγRIIB and complement-mediated process for subsequent presentation [48] which seems more likely to be a requirement for prion pathogenesis under the current prion hypothesis. Uptake of prion agent by other Fc receptors may mediate intracellular processing, e.g., for presentation on MHC class II molecules. Extreme processing of the prion agent is likely to render it ineffective. Cleavage of the cellular prion protein has been shown to protect against infection [49], whilst cleavage of PrP^Sc has been shown to modulate prion propagation in a similar fashion [50].

Complement component C1q also regulates dendritic cell development from monocytes via DC-SIGN signaling [51,52]. This induces more tolerogenic DCs and alters their cytokine production, leading to increased Il-10 and reduced Il-12 production and increased phagocytosis of apoptotic cells [53,54]. These data would suggest that C1q deficient mice may also possess a bias in their DC component or functions, as well as a deficiency in immune-complex trapping to FDC, further complicating the interpretation of prion pathogenesis studies performed within them.

The role of complement during prion pathogenesis is further complicated by the observation that the individual complement components mediating interaction with PrP^Sc appear to differ dependent upon the strain of infectious agent [55]. Cell-autonomous and complement-assisted cell-mediated waves of prion trafficking occurs to lymphoid follicles, the relevance of each transport mechanism to disease pathogenesis remains undetermined [56].

The role of the complement system within the CNS has also been extensively reviewed recently [57,58]. As with most innate immune functions within the CNS a clear neuroprotective/neurodegenerative dichotomy exists within the literature, likely indicative that complement activation may lead to both protective and degenerative effects dependent upon the context, responsive and regulatory mechanisms involved. There is currently little evidence supporting a role for complement in prion pathogenesis within the CNS. Mice lacking C1qa, C2 or C3 revealed no deficit to prion pathogenesis following intracerebral inoculation [44] and mice lacking both complement components C3 and C4 revealed unaltered pathogenesis following aerosol exposure to prions [11].

Mast cells have been implicated in prion pathogenesis due to their high expression levels of cellular prion protein (PrP or PrP^C) and their ability to traffic to the brain [59]. In fact mast cells are resident, and are also capable of trafficking from the blood to specific regions within the CNS. These different populations of mast cells differentially express the c-kit receptor [60]. Mast cells function to release histamine (via binding of allergens to FcεR1α) and heparin in an early response to infection, and are critical to induce chemotaxis and infiltration of neutrophils and MNPs. Due to the non-inflammatory nature of prion infection there is little evidence to suggest that mast cells undertake this conventional role during peripheral prion pathogenesis. Studies of the activation pathways in mast cells reveal that other factors such as C3a and CCL3 may trigger mast cell degranulation without IgE binding to FcεR1α [61]. This activation pathway is most likely to occur during prion pathogenesis as there is little evidence of the induction of an adaptive immune response and specific anti-prion antibody generation during pathogenesis. The role of immediate local responses to prion infection is in clear need of further investigation to determine the activation effects on mast cells, MNPs and neutrophils amongst other innate immune cell types.

The role of mast cells within the brain is thought to be neuro-modulatory and mast cell trafficking to the brain is linked to steroidal hormones and sexual activity or anxiety behaviors [62]. The presence of mast cells within the brain and their ability to shed expressed PrP upon activation [59] may have implications for prion pathogenesis within the CNS. Shedding of PrP^C by mast cells likely occurs via proteolytic or lipolytic cleavage mechanisms removing the glycosylphosphatidylinisotol (GPI) anchor from the protein [59,63]. Activation of mast cells within the CNS may provide copious extracellular PrP^C substrate for conversion to the disease-associated isoform late in the disease process, when significant damage has occurred to the CNS already, adding to the exponential accumulation of misfolded protein within the clinical stage of disease. Investigation of GPI-anchorless PrP transgenic mice has revealed that they are capable of forming amyloid in the brain without clinical prion disease symptoms, whilst co-expression of anchorless and native PrP^C increased prion pathogenesis [64]. Novel synthetic prion strains have subsequently been generated in the above mentioned co-expressing model [65], suggesting that mast cell involvement during infection may aid in or promote the natural prion strain mutation [66] or prion evolution phenomena [67].

MNP constitute a system of highly phagocytic cells that have a common origin in the bone marrow and circulate or reside in the body tissues in order to sample their environment via uptake of foreign (and self) material. The development of MNP relies upon expression of the colony stimulating factor 1 receptor (CSF-1R) during their differentiation in order to respond to CSF-1. MNP play a vital role in the bridge between the innate and adaptive immune systems by processing and presenting of antigenic compounds to stimulate specific antibody production. During the initial processing by MNP material may be directly disposed of by degradation via lysosomal or proteasomal pathways. These actions of MNP aid in the development of immunity or tolerance to the vast multitude of foreign material encountered. Numerous (particularly intracellular) pathogens have adapted to abuse the functions of MNP as a vehicle for pathogenesis, allowing for their uptake into MNP and trafficking to lymphoid organs for further spread. It is most likely that the uptake and spread of prions to lymphoid organs occurs using MNP as a ‘Trojan horse’ via a similar mechanism.

The role of MNP in prion pathogenesis have been subject to recent review and are considered diverse [68]. This reflects the wide diversity of MNP and their roles in innate immunity. To simplify this diversity MNP can be broadly categorized into the following: (1), resident cells with degradative functions; (2) resident cells with antigen presenting functions and (3) systemic circulating cells responsive to inflammatory stimuli. At present there is little evidence implicating circulating ‘inflammatory monocyte’ populations in prion pathogenesis due to the non-inflammatory nature of infection.

Regardless of their major functionality all MNP likely use similar surface receptor based pathways and mechanisms to sense their local micro-environment and tailor their responses to pathogens or tissue damage. The exact receptor-mediated mechanism determining the response to prion agents has not been identified or definitively characterized. MNP-mediated uptake of prion agent is enhanced by complement opsonization (see Section 2.1 above). The uptake of prions likely involves complement, lectin or scavenger receptors while there is evidence that Fc [44] (FcγR, RII and RIII deficient mice reveal no deficit following i.c. or i.p. inoculation), or toll-like receptors [69] (at least via the Myd88-dependent pathway) have little role in peripheral pathogenesis under steady state conditions. The downstream processing of the prion agent will likely depend upon the MNP type it has been encountered by.

Gene expression data reveal that PrP expression is generally down-regulated during granulocyte differentiation [151] though some expression is still detectable [152] (see also Figure 2). Neutrophil functions have been shown to be inhibited by both native and scrapie associated prion protein, resulting in failure of neutrophil aggregation and deficits in superoxide radical and beta-glucuronidase export [153]. A 20 amino acid fragment of the prion protein sequence (termed PrP^106-126) has been shown to be directly neurotoxic, activate MNPs and act as a chemotactic agonist for the FPRL1 receptor [121]; the relevance of the PrP^106-126 fragment to prion pathogenesis is elusive at best. FPRL1 is strongly expressed on granulocytes and MNP cell types and promotes chemotaxis to sites in response to allergic inflammation [154]. Eosinophilic inclusions have been observed in both central and peripheral amyloid deposits, although amyloid deposits themselves are eosinophilic, there is some evidence to suggest eosiniophils are present within or around peripheral amyloid deposits. The basis for this localization appears to lie in FPRL1 chemotaxis.

Natural killer (NK) and γδ T cells are thought to play little role in the pathogenesis of prion diseases. Both cell types require triggering or activation by danger signals, for NK cell this is usually via cytokines and for γδ T cells little is known about their activating signals. NK cells primarily respond to virally-infected cells and function to destroy these cells via expression of perforin and granzyme. Perforin-knockout mice revealed unaltered prion pathogenesis [88]. NK cells are typically stimulated via interleukins, 2, 12, 15 and 18 and chemokine CCL5 as well as down-regulation of MHC class I molecules following viral infection. The lack of reported NK cell response to prion pathogenesis suggests that prion agent uptake by monocytes may fail to activate NK cells. There are numerous viral mechanisms used to evade NK cell activation including regulating apoptosis, modulating cytokines and chemokines, and compromising DC functions for review see [155,156,157]. Little evidence has been discovered for any of these functions following prion infection apart from the possible induction of apoptotic cell death mechanisms (in neurons). Recently NK/T cells (a subset of T-cells that express both the NK1.1 marker and αβ T cell receptor) in the spleen have been associated with prion infectivity [91]. Investigating the role of NK/T cells in prion pathogenesis is hampered by the fact that though Rag2^−/− mice are deficient in NK/T cells they are unsusceptible to prion infection due to their B-cell deficiency and failure to generate mature FDC. Selective reconstitution of the Rag2^−/− model for prion pathogenesis studies has not been reported. An increase in γδ T cells in the peripheral blood mononuclear cells (PBMCs) has been reported following scrapie infection in sheep [116].

The megakaryocyte lineage and platelets have been associated with the expression of prion protein [123,158] and MicroRNA miR146a [159]. Following activation of platelets PrP^C is released in exosomes [160], similar to the release of exosomal PrP^C from macrophages [161]. While not classically considered part of the innate immune system, evidence regarding the expression of TLR2 having a regulatory function has suggested a new role for megakaryocytes in immunity [162]. The over-expression of miR146a, TLR2 and TLR4 has also been reported in microglial cells during scrapie (prion) infection [139]. These data may indicate some common co-expression mechanism of these genes or their possible functional interaction. Regardless of their role in immunity megakaryocytes and platelets constitute a significant portion of the PrP expressing bodies within the blood compartment and as such represent a major vector for prion transmission during blood transfusion [163,164,165,166]. Spread of prion infection between lymphoid organs (i.e., from draining to contralateral non-draining lymph nodes) has been prevented by blocking the egress and thus recirculation of B-cells [143]. The routing of prion pathogenesis has been characterized in precise immunohistological detail and describes neuroinvasion from lymphoid organs via the peripheral nervous system [167,168]. Hematogenous spread from lymphoid organs to the CNS has been proposed as an alternative or complementary pathway to neuroinvasion in natural ruminant prion disease with involvement of the circumventricular organs (CVO) in the CNS [169].

Expression of the cellular prion protein (PrP^C) by the hematopoietic compartment is not required for prion pathogenesis [170], confirming that the innate immune system function in disease pathogenesis operates via non-PrP^C dependent mechanisms. Little is known about the uptake mechanisms of prions and the factors that lead to cell clearance, infection or passivity that may facilitate transport. PrP^C is not required for uptake of prions in vitro, even by non-phagocytic cell types such as neurons [171]. There has been much debate regarding which cells of the immune system do or do not express appreciable levels of PrP. It has long been known that PrP is strongly expressed on hematopoietic stem cells within the bone-marrow compartment; specific lineages appear to lose or down-regulate this expression during differentiation [172,173]. Maturation of cells in response to all-trans retinoic acid is one mechanism known to down-regulate PrP expression [151]. In light of the current debate regarding hematopoietic differentiation and retention or loss of PrP^C expression we present here data on the expression of the prion protein gene (Prnp) transcripts in various innate immunity relevant cellular compartments (Figure 2). The ability to detect Prnp gene expression and cellular prion protein differ, indeed it appears that the mature protein is often more labile or difficult to detect than its corresponding mRNA message, possibly accounting for some of the reported discrepancies [174,175]. The data we present here clearly depicts altered Prnp expression levels between subcomponents of the hematopoietic system with Mϕ, DC, Microglia, LC and IFN-producing killer DC displaying the highest observable levels of expressed Prnp transcript.

The most common mechanism for determining or proving a correlation between components of the innate immune system (be they cells or proteins) and prion pathogenesis is via transgenic mouse models. These models have been used to determine the effect of knockout or overexpression of a particular protein or cell type on prion pathogenesis. The classic experiment revealed that knockout of Prnp resulted in complete resistance to prion infection [140,141]. A plethora of immunity-associated candidate genes have been identified by observational techniques e.g., following gene expression profiling and numerous genes have been screened for a role in prion pathogenesis via knockout transgenic mouse models, e.g., [130] (see also Table 1). The majority of such studies have determined that knockout of an individual component does not prevent prion pathogenesis, for some a slight alteration in pathogenesis was observed and for relatively few a block of pathogenesis was observed and then often only under certain specific conditions. We have performed a pathway analysis of the innate immunity associated genes implicated in prion pathogenesis and determined the major host upstream regulators of prion pathogenesis-associated gene expression (Table 2).

From these candidates all have previously been implicated with prion pathogenesis, Il-4, Il-6 and Il-12 have been shown not to be required. Il-10 knockout mice revealed major alterations to, but not prevention of, prion pathogenesis. These data reveal the cytokine milieu occurring during prion pathogenesis and may underlie the basis of alternative activation of MNP and lack of inflammatory, or tolerization of, responses during prion infection. These upstream regulators accounted for 118 out of the 144 genes analyzed presented as a wheel network diagram to reveal the overlap between the regulators and target genes (Figure 3), perhaps revealing why knockout of any individual component may have little overall effect upon prion pathogenesis. Many of these genes and their regulators effect MNP differentiation, maturation and homeostasis as well as modulating MNP responses during infection. These findings indicate that prion pathogenesis is influenced by the steady-state, activation and response of the innate immune system.